Research On Wheel Tread Wear Prediction And Profile Optimization Of High-speed Train Based On Identity Mapping Multi-layer Extreme Learning Machine

China ’s high-speed railway network has been extended to a variety of complex road conditions.As one of the main travel modes of Chinese residents,high-speed trains are constantly increasing.With the increase of train operation time,the changing service environment of high-speed trains leads to the problem of wheel-rail wear.At the same time,during high-speed operation,even small wheel-rail wear and rail excitation will pose a serious threat to the safety of trains.Also,the complex natural environment increases the difficulty of wheel-rail wear analysis to a certain extent.Therefore,this thesis deeply investigates the research status of wheel-rail relationship,wheel-rail wear prediction,extreme learning machine and wheel-rail profile optimization of railway vehicles,and carries out research on wheel tread wear prediction and profile optimization of high-speed trains based on swarm intelligence algorithm and neural network.The specific research work of this thesis is as follows:(1)The basic principle of identity mapping is introduced into the multi-layer extreme learning machine,and the identity multi-layer extreme learning machine model(I-ML-ELM)is established.The network performance of the model is analyzed by different types and different dimensions of public data sets and different types of network models.The results show that the I-ML-ELM model established in this thesis shows better prediction performance and better generalization performance than other networks in the public data set,and can show stronger prediction applicability in different data sets,and has higher operating efficiency.Finally,the control variable method is used to test the influence of different neuron numbers on the network prediction performance,which lays a foundation for the application of the later network.(2)Based on the vehicle-track coupling dynamics theory,the vehicle-track coupling dynamics model of high-speed train single trailer is established.The Kalker linear theory and Archard wear theory based on friction work are used to simulate and analyze the wheel wear under different working conditions.The influencing factors of wheel wear and the evolution law of wheel wear of high-speed trains are analyzed.The research on wheel tread wear prediction based on identity mapping multi-layer extreme learning machine is carried out.By changing the operating conditions of the train,the wheel wear data under different working conditions are obtained,which is used as the data set to train I-ML-ELM,and the wheel tread wear prediction model based on I-ML-ELM is established.The numerical results show that the wear prediction model established in this thesis can effectively predict the maximum wear value of wheel tread,and has higher computational efficiency than other networks.(3)By introducing chaotic mapping and random perturbation,the golden jackal optimization algorithm is improved in the early and late stages of optimization calculation,and a golden jackal optimization model based on random perturbation Sine-Tent-Cosine chaos(RPSTC-GJO)is established.The performance of the model and other optimization algorithms is tested using a variety of different types of benchmark functions.The test results show that RPSTC-GJO shows better convergence characteristics and faster convergence speed than ASO,PSO and GJO algorithms in different benchmark functions.(4)Combined with RPSTC-GJO optimization model and I-ML-ELM prediction model,the optimization design of high-speed train wheel profile is carried out.Taking RPSTC-GJO algorithm as the main structure,a multi-layer extreme learning machine model based on identity mapping is embedded to realize the mapping between solution space and target space.Finally,the optimal solution meeting the target requirements is obtained through iterative optimization.The dynamic analysis and calculation of the high-speed train trailer model with the optimized profile are carried out.The results show that the optimized profile can reduce the wheel-rail wear to a certain extent under the condition of ensuring the stability,stability and safety of the train.